Even though there was early interest in the neural basis of automatic behavior in the beginning of the previous century, it is still unclear how we obtain and express habits.The dominant view of the 20th century was that novel behaviors are generated by cortex, because they require attention and flexible thinking. Contrary, habits are automatic, so they are primarily mediated by subcortical structures, more precise basal ganglia (BG). Nowadays, there is arising bibliography that argues exactly the opposite position. So the essential question is "Who's the teacher and who's the pupil?".Though our computational model of cortical-basal-thalamic closed loop, we show how learning has been transferred from the basal ganglia to the cortex, simply as a consequence of the statistics of the choice. Our theory was also tested on monkeys, which confirmed our predicted results.